Amazon’s U.S. search is widening the shop window while other stores handle the sale. That deserves attention: owning checkout does not mean owning customer discovery. Google’s expanded AI search ads raise a related question about advertisers’ control as distribution grows. eBay adds a financial check: more marketplace sales can coexist with narrower margins. The distinction worth watching is reach versus results.
News
1️⃣ Amazon expands outside-store product discovery in the U.S.
2️⃣ Google expands AI search ads, with limited separate reporting.
3️⃣ eBay marketplace sales rise while operating margins narrow.
Insights
4️⃣ Italy’s marketplace map calls for category-specific channel choices.
5️⃣ EU AI use rises faster in large enterprises than small ones.
6️⃣ An NP Digital visual questions search traffic forecasting assumptions.
LET’S ENTER THE MAZE!
1️⃣ News
The Maze: Amazon’s U.S. shopping service reportedly now lists 250 million products sold by other stores. The expansion makes Amazon a broader starting point, even when the sale happens elsewhere.
Source: Amazon. Frame from Amazon’s March 2026 demonstration showing outside-store sneaker listings with Shop direct and Buy for Me options.
The Details:
The catalog has grown. Shop Direct’s 250 million listings are 2.5 times March’s level. Amazon is also making these offers more prominent and showing more product details.
Checkout has two routes. Shoppers can visit the merchant’s website. For eligible items, Buy for Me lets them authorize Amazon to complete that website’s purchase. The merchant handles delivery, returns and service.
Search exposure is not sales. Amazon says outside-store products appear in hundreds of millions of searches monthly. The expansion report provides no completed-order or merchant-profit measure.
MarketMaze Take: Amazon can send a shopper to another store and still remain the place they begin. A merchant’s own checkout does not guarantee independence from discovery platforms.
The lesson travels. Merchants elsewhere can retain checkout yet depend on another company to introduce customers. That is the broader issue; this service’s documented availability is U.S.-only.
Demand needs proof. More listings create potential exposure. Without incremental orders or profit data, the larger catalog proves reach rather than merchant success.
2️⃣ News
The Maze: Google is expanding ads in AI Mode, its conversational search experience, to ten more countries. Existing advertisers gain placements, with limited ability to separate their results.
Source: Google. Google’s May 2025 demonstration shows a Sponsored website-builder result inside an AI Mode answer; it does not document the October 2026 rollout.
The Details:
Access has boundaries. English-language ads are scaling in Australia, Canada, India, Kenya, Malaysia, New Zealand, Nigeria, Pakistan, the Philippines and Singapore. Pakistan gets text ads only. Brazil, Mexico and the UK remain planned for Q4 2026.
Budgets already participate. Eligible Search, Shopping and automated Performance Max campaigns enter relevant auctions without separate campaigns or budgets. Ads can match a product query or the commercial context of an answer. Follow-up-question ads launch separately in the US.
Evidence stops at clicks. Google compares how often sponsored and unpaid product-list items get clicked, using internal US data. It discloses no sales lift. AI Mode results enter ordinary campaign reports in aggregate.
MarketMaze Take: Easier distribution is useful. It is not proof of better economics. Google is extending where existing budgets compete before providing a distinct channel businesses can evaluate separately.
Control matters beyond these markets. Advertisers cannot target only AI Mode or opt out separately from Search. The wider lesson is platform dependence, not worldwide availability.
Judge customer outcomes. Existing keyword exclusions remain available. More placements may help, but similar click rates do not establish cheaper customers or additional sales.
3️⃣ News
The Maze: eBay’s global marketplaces handled US$23.2 billion in third-quarter transactions, up 15% from a year earlier. Its own revenue rose 14%, but operating margins narrowed.
The Details:
Sales are not revenue. Transaction volume includes shipping and taxes, before returns or cancellations. eBay’s US$3.206 billion revenue comes from fees, advertising and services.
Acquisitions enlarge the base. Fashion resale marketplace Depop joined eBay on July 30. Active buyer accounts grew 10%; excluding Depop and acquired resale platform Tise, they grew 1%. Buyers are measured over 12 months.
Promotion grew; margins fell. eBay’s first-party advertising revenue rose 16% to US$575 million. Operating profit increased in dollars, but its share of revenue fell to 19.2% from 20.3%.
MarketMaze Take: A larger platform does not automatically mean better seller economics. The sales headline needs a local demand check and a separate profit calculation.
Location matters. U.S. transaction volume grew 27%; international volume grew 4%. Sellers elsewhere cannot use the global growth rate as their local demand forecast.
Promotion needs its own return. Advertising revenue is money eBay earns, not sellers’ profit. Acquisitions complicate the sales comparison, but do not establish that underlying demand stalled.
4️⃣ Insight
The Maze: Italy's marketplace landscape is crowded in specialists, local C2C and category retailers. The unresolved gap sits with grocery-led generalist retail.
Lengow places Amazon, eBay, AliExpress, Temu, TikTok Shop, Subito and Bakeca.it in the same generalist pure-player space.
Specialist depth is the market's real signal: fashion, refurbished tech, DIY, flash sales, parapharmacy, food, books, sports, beauty and home all have visible players.
The open question is grocery, where Conad, Coop, Esselunga, Selex and Eurospin have reach and trust but no clear open marketplace model.
Why it matters: Italy is not a one-platform expansion plan. Brands need category-specific marketplace choices, while grocery remains the strategic gap.
5️⃣ Insight
The Maze: AI adoption is rising across Europe, but not evenly: large enterprises hit 55.03% usage in 2025, while small firms reached only 17.00%.
EU enterprises with 10+ employees rose from 13.48% AI usage in 2024 to 19.95% in 2025 in Eurostat's dataset.
Large firms moved 13.86 percentage points in one year, while small firms moved 5.79 points, widening the size gap.
The small-firm opportunity is real, but the data says process capacity still beats tool access.
Why it matters: AI is cheaper than old enterprise software. Implementation is not. SMEs need one painful workflow, one owner, and one clear payoff.
6️⃣ Insight
The Maze: AI search is not only stealing clicks. It is breaking the old bridge between ranking, sessions, and demand forecasts, so neat traffic plans get noisy fast.
NP Digital's visual shows AI Overview / zero-click pressure rising from 59% in 2024 to 72% in 2026, while forecast accuracy falls from 55% to 48%.
The 2026 spread is the warning: 72% zero-click pressure versus 48% forecast accuracy, a 24-point gap between search behavior and planning confidence.
Support research shows AI Overview sources can differ from classic Google results, so rank tracking no longer explains the full influence surface.
Why it matters: Ecommerce teams need forecasts that track answer visibility, citation share, branded demand, and click leakage. Rank is still useful. It is no longer enough.
🧰 Resources
Useful stuff, handpicked by the team every day
🤖 AI prompts: 12 AI Prompts for Ecommerce Dashboards & Reporting or 👉 All here
📖 Glossary: Meta Ads, Decoded: 16 Terms You Actually Need to Know or 👉 All here
📅 Events: DELIVER America 2026 or 👉 All here
🔓 Get the full resource library with MarketMaze+
🗞️ Quick hits
Everything else you should know
🛍️ Shopping and marketplaces
Shopify adds one checkout across eligible Shop app stores and Shop Pay in personal AI agents. Buyers can pay once across stores, but shipping charges and orders stay separate.
Etsy’s Q3 marketplace sales rose 8.8% to $2.6bn, and annual active buyers grew 1% to 88m for the first time in two years. Spending per buyer rose 3.9% over twelve months, while purchase frequency was still lower.
🔎 Advertising and search
Meta says Facebook and Instagram ads can learn from similar products before building their own performance history. Advertisers need no new setting; the system still has to learn how people respond.
Google says query growth is its fastest in five years, while annual Search revenue growth slowed to 12%. Currency, spending trends and lower AI Mode monetisation complicate the comparison; merchant returns remain unmeasured.
💳 Payments and tax
Wero starts replacing iDEAL in the Netherlands: eleven banks are ready, but shops migrate through their payment providers toward an end-2027 goal. Dutch purchase protection follows its own rollout.
The UK plans to make marketplaces collect value added tax on qualifying UK-business sales from October 2027. Personal and second-hand sales are excluded; compliant retailers still await refund details.
🚢 Freight
C.H. Robinson says its AI has produced over 80,000 ocean and air freight quotes for 3,600-plus customers, with quotes in two to three minutes. Its separate worldwide ocean-booking rollout is still an early 2027 target.
THAT’S IT FOR TODAY!
You’re the reason our team spends hundreds of hours every week researching and writing this email.
What do you think of this issue?
See you next time in the maze!
MarketMaze team
